The bug was discovered by compiling and testing Labkit against imglib2-7.1. Trying to use the brush in Labkit would trigger this NullPointerException.
I'm unsure about the rules for implementing the new getType() so please have a closer look at my suggested change and whether is correct.
I also noticed that (A)imgLabeling.getType().createVariable() would behave differently than (B)imgLabeling.randomAccess().get().createVariable(). (A) creates a LabelingType with a independent LabelMapping while (B) creates a LabelingType with the same LabelMapping as the imgLabeling. In Labkit I'm now using implementation (B) because a need the same LabelMapping in order for LabelingType.set(...) to work properly.
The bug was discovered by compiling and testing Labkit against imglib2-7.1. Trying to use the brush in Labkit would trigger this NullPointerException.
I'm unsure about the rules for implementing the new getType() so please have a closer look at my suggested change and whether is correct.
I also noticed that (A)
imgLabeling.getType().createVariable()
would behave differently than (B)imgLabeling.randomAccess().get().createVariable()
. (A) creates a LabelingType with a independent LabelMapping while (B) creates a LabelingType with the same LabelMapping as the imgLabeling. In Labkit I'm now using implementation (B) because a need the same LabelMapping in order for LabelingType.set(...) to work properly.